ctx engineering, inc.
ctx engineering, inc. develops open-source tools for context engineering: software that helps coding agents search and retrieve their local session history, attribute code to the agent session that produced it, map code relationships, and reduce noisy tool output. Its projects are designed to run locally by default, with optional sharing and integrations.
At a Glance
- Individual developers using coding agents
- Software teams coordinating multiple coding-agent sessions
- Developers and organizations training coding models
- Teams working in large codebases that need graph-based navigation
- +1 more
AI Tools by ctx engineering, inc.
(2)figma-server
Figma MCP Server for AI Agents
ctx
Multi Agent Dev Environment
Discussions
No discussions yet
Be the first to start a discussion about ctx engineering, inc.
Latest News
ctx v2.2.8 released with incremental Codex import and Blame-maintenance fixes
ctx v2.2.7 released with more efficient Codex imports and Sift telemetry integration
figma-server v0.2.0 prerelease published; Apache-2.0 browser/CDP access for Figma agents
ctx 2.2.0 introduced an opt-in hosted-history beta for invited team sharing and administration
Products & Services
Open-source Rust CLI/toolchain that indexes local coding-agent histories and lets users search, inspect, and locate complete sessions and events; it also includes Blame, Graph, and Sift capabilities in the unified 2.x executable.
Code-attribution feature that traces a file, line range, commit, or pull request to the agent session that produced it, with citations to transcript and tool-call evidence.
Rust codebase and documentation graph that stores an indexed graph in SQLite and supports symbol search, callers/callees, impact queries, paths, reports, and exports.
Local tool-output compactor that removes repetitive paths, logs, and formatting while retaining useful details; optional original-output retention enables recall.
Market Position
ctx positions its unified local toolchain against raw transcript search, agent-memory/summary tools, and Graphify: it emphasizes exact searchable records and evidence rather than lossy summaries, local-first privacy, and a Rust implementation. ctx graph is described as an independent, faster rebuild of Graphify rather than a fork or drop-in replacement; the site reports 1,000x faster graph search and 89x faster updates in its benchmark. Sift is positioned against RTK-style output compaction, while figma-server differentiates from Figma's access-controlled native MCP server by driving the Figma web app through Chromium/CDP.
Leadership
Founders
Luca King
Founder and Software Engineer at ctx.rs. Previously Founder at Profound Health Institute. Background in software engineering and student of Royal Holloway, University of London.
Executive Team
Luca King
Founder
Creator of the ctx Agentic Development Environment and proponent of the 'FermAI Paradox' concept.
Founding Story
ctx was created to solve the 'FermAI Paradox'—the disparity between what coding agents generate and what they accomplish in production. The goal was to unify fragmented agent workflows into a single interface with durable transcripts and secure runtimes.
Business Model
Revenue Model
The core projects are published as open-source software and can be installed from GitHub releases, Homebrew/mise, or source. The public materials describe an optional hosted-history/server beta and support channels, but do not state a paid pricing model.
Target Markets
- Individual developers using coding agents
- Software teams coordinating multiple coding-agent sessions
- Developers and organizations training coding models
- Teams working in large codebases that need graph-based navigation
- Design and product teams whose agents need to inspect or edit Figma files
- Recover decisions, constraints, failed investigations, and assumptions from earlier coding-agent sessions
- Avoid repeating work across parallel or resumed agent threads
- Audit or explain code by tracing it back to the session that wrote it
- Navigate large codebases and assess symbol impact or dependencies
- Reduce token consumption from repetitive command output
- Train coding models and provide agents with searchable local history